PGC-1α supports glutamine metabolism in breast cancer

1Altered metabolism is an emerging hallmark of cancer cells. The most studied metabolic alteration of cancer cells is their high reliance on glycolysis even if there is sufficient oxygen to support mitochondrial oxidative phosphorylation (Warburg effect). The rapid ATP production of glycolysis is thought to benefit the enhanced proliferation of cancer cells. However, glutamine metabolism has been increasingly recognized as a central metabolic pathway in cancer, and it is thought that the catabolic and anabolic roles of glucose and glutamine cooperate in fuelling tumor growth. The discovery that the oncogene Myc is an important regulator of glutamine metabolism has further strengthened the importance of this metabolic pathway in cancer.

2Recent studies revealed that reductive carboxylation of glutamine to citrate through reversal of the citric acid cycle (CAC) and the subsequent use of citrate for lipogenesis constitute a key anabolic route in hypoxia in cancer cells, as well as in cancer cells with mitochondrial dysfunctions. Hypoxia-induced factor 1 (HIF-1) plays an important role in this metabolic reprogramming, as cells constitutively expressing HIF-1α perform reductive carboxylation of glutamine in normoxia. Clearly, the ability to rely on this metabolic pathway will be a distinct advantage for solid tumors that experience fluctuating levels of oxygen by permitting elevated production of lipids, which are important notably for membrane biogenesis.

3The family of peroxisome proliferator activated receptor gamma coactivator 1 (PGC-1) consists of three members, PGC-1α, PGC-1β, and PRC. These transcriptional coactivators play a central role in metabolism. Even though the importance of the PGC-1s in regulating mitochondrial functions and metabolism is well established in general, their roles in regulating cancer metabolism and proliferation have only started to emerge recently. The PGC-1s have been shown to have both pro- and anti-tumorigenesis effects. In breast cancer, PGC-1α has been shown to promote the growth of ERBB2+ tumors in vivo and to regulate their mitochondrial metabolism and angiogenic properties. In this paper, we inquired whether PGC-1α participates in the control of glutamine metabolism in breast cancer. We reveal that PGC-1α increases the expression of glutamine metabolism genes, augments glutamine-mediated forward and reverse CAC fluxes, as well as elevates glutamine-mediated lipogenesis in hypoxia. This PGC-1α-mediated increase in glutamine metabolism will confer an advantage to solid breast cancer tumors experiencing limited nutrients and oxygen supply. Relevance to the human disease is underlined by the findings that PGC-1α expression is positively correlated with that of the glutamine pathway in ERBB2+ breast cancer patients, and that high expression of the glutamine pathway is associated with shorter survival time.

4All procedures were conducted in accordance with approved animal protocols by the McGill University Animal Care Committee. Wild-type mice (C57BL/6J) and muscle creatine kinase (mck)-PGC-1α transgenic mice (C57BL/6-Tg(Ckm-Ppargc1a)31Brsp/J) as described in were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). All animals used were female and were sacrificed at approximately 12 weeks of age.

5Stable cell lines Control-1, PGC-1α-1.1, and PGC-1α-1.2 have been described in. These cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM), 10% FBS, 10 μg/mL insulin, 20 mM HEPES, penicillin/streptomycin, 1 μg/mL puromycin, 400 μg/mL G418, at 37°C and 5% CO2. SK-BR-3 and BT-474 were obtained from the American Type Culture Collection (ATCC) and cultured in DMEM supplemented with 10% FBS and penicillin/streptomycin. Hypoxia experiments were performed in a Thermo Scientific (Rockford, IL, USA) Heraeus HERAcell 150 set at 1% O2 and 5% CO2. Pharmacological inhibition of ERRα was performed using Compound 29 (C29) at 5 μM (0.1% DMSO). C29 was synthesized according to.

6Total RNA from cultured cells, grown to 50% to 60% confluence in 35 mm plates, was extracted using the Aurum Total RNA Mini Kit (Bio-Rad, Mississauga, Canada) and was reverse transcribed with iScript cDNA Synthesis kit (Bio-Rad). mRNA expression analyses by real-time PCR were performed using iQ SYBR Green Supermix (Bio-Rad) and gene-specific primers with the MyiQ2 Real-Time Detection System (Bio-Rad). Values were normalized to TATA binding protein (Tbp) for murine cell lines or beta-2 microglobulin (B2M) for human cell lines.

7Chromatin was prepared from PGC-1α-1.1 cells, using approximately 107 cells per chromatin immunoprecipitation (ChIP). ChIP was performed as described previously using a specific anti-PGC-1α antibody (Santa Cruz Biotechnology, Inc., Dallas, TX, USA, catalog #sc-13067) or specific anti-ERRα antibody (Epitomics, Burlingame, CA, USA, catalog #2131-1). Quantification of PGC-1α or ERRα ChIP enrichment was performed using real-time PCR using the LightCycler 480 instrument (Roche, Mississauga, Canada) and specific primers for genomic regions. ChIP enrichment was normalized against the control region and further normalized against the IgG control.

8Total proteins from cultured cells or muscle tissue were extracted with lysis buffer (50 mM Tris–HCl pH 7.4, 1% Triton X-100, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA) with inhibitors (2 μg/mL pepstatin, 1 μg/mL aprotinin, 1 μg/mL leupeptin, 0.2 mM phenylmethylsulfonyl fluoride and 1 mM sodium orthovanadate) and quantified with the Bio-Rad Protein Assay kit (Catalog #500-0006). The blots were incubated according to the manufacturer's instructions with the following primary antibodies: Glud1 (GeneTex, Irvine, CA, USA, GTX88164), Gls (Abcam, Toronto, Canada, AB93434), Got1 (GeneTex, GTX88903), Got2 (GeneTex, GTX88925), Hsp90 (Santa Cruz Biotechnology, sc-101494), and Actin (Santa Cruz Biotechnology, sc-1616) and with horseradish peroxidase-conjugated secondary antibodies (GE Healthcare, Mississauga, Canada, or Santa Cruz Biotechnology). The results were visualized using Western Lightning Plus-ECL (Perkin Elmer, Waltham, MA, USA) or SuperSignal West Femto ECL (Pierce, Rockford, IL, USA). Densitometry analyses were conducted using ImageJ software (NIH).

9BT-474 cells were subjected to either 10 nM control siRNA (Dharmacon, Rockford, IL, USA, D-001810-10-05) or a combined 10 nM pool of four siRNA specifically targeting PPARGC1A (Qiagen, Germantown, MD, USA, FlexiTube GeneSolution GS10891). Cells were transfected using calcium phosphate (CaPO4) and incubated for 120 h before extraction.

10Cells were seeded at a density of 400,000 cells per 35 mm plate in standard growth media and conditions. After 24 h, the medium was supplemented with trace amounts of [U-14C]-glutamine (0.094 μCi/mL) and the cells were further incubated either in normoxia (21% O2) or hypoxia (1% O2) for 24 h (Control-1 and PGC-1α-1.1) or 48 h (BT-474). Individual experiments were conducted in triplicate with additional plates for cell counting. Cells were washed with PBS, collected using 0.5% Triton X-100, and stored at −80°C. After thawing, lipids were extracted using chloroform/methanol; the lipid-rich chloroform phase was dried under air and resuspended in 150 μL chloroform. Scintillation counts were measured over 5 minutes per sample using a MicroBeta2 scintillation counter (Perkin-Elmer) and Eco-Lume reagent (MPBio, Solon, OH, USA). Counts were normalized for cell number.

11Cellular glutamine uptake was measured using the BioProfile 400 analyzer (Nova Biomedical Corp., Waltham, MA, USA). Following growth under normoxic (21% O2) or hypoxic (1% O2) conditions, media were removed and maintained on ice until analysis. Glutamine uptake was calculated as the difference in glutamine content between culture media and unseeded media incubated in parallel plates and normalized for cell count.

12The metabolomics analyses by GC/MS were adapted from. Additional details are provided in Additional file 1.

13Breast cancer data, clinical data, and survival data for 732 patients were obtained from The Cancer Genome Atlas (TCGA). RNAseq data were downloaded from TCGA web portal using the reads per kilobase of exon model per million mapped reads (RPKM) quantification. RPKM data was log2 transformed and expression values lower than 0.25 RPKM were removed. To generate an overall pathway score, all nine genes in the glutamine metabolism pathway were Z-transformed (mean 0 and standard deviation of 1). Mean of zero corrects for the fact that the genes may be expressed at different levels. Transforming the standard deviation to 1 causes the genes to change in the same magnitude, so we can see if the overall pathway is over/underexpressed. If the genes are expressed at different levels, the changes in the pathway will be dominated by the genes that were originally expressed at higher levels. We used the median of Z-scores to generate a pathway score, which was used in correlation and survival analysis.

14All statistical analyses were performed using GraphPad Prism (GraphPad Software, Inc., CA, USA), Microsoft Excel (Microsoft Corporation, CA, USA) or Matlab (The Mathworks, Inc., MA, USA). Survival analysis was calculated using an implementation of Kaplan-Meir log rank testing from Matlab Exchange.

15Additional methods are provided in Additional file 1 and in Additional file 2: Table S1.

16To assess the impact of PGC-1α on the expression of glutamine metabolism genes, we used clones of ERBB2/Neu-induced breast cancer cells with modestly increased expression of PGC-1α (α-1.1 and α-1.2) and control cells (Ctl-1). These cells have been extensively characterized, and as expected, they display an approximately 30% increase in expression of well-known PGC-1α mitochondrial target genes [see Additional file 3: Figure S1A]. Here we reveal that the mRNA expression of enzymes involved in glutamine catabolism, transport, and synthesis were significantly increased in PGC-1α clones compared with control (Figure 1A,B). In support of this point, cells with reduced expression of PGC-1α displayed lowered expression of glutamine metabolism genes [see Additional file 3: Figure S1B]. In addition, PGC-1α controls these genes in a cancer-independent context, as PGC-1α muscle-specific transgenic animals exhibited increased expression of several glutamine metabolism genes [see Additional file 3: Figure S1C,D]. Given the known functional relationship between PGC-1α and the estrogen-related receptor α (ERRα) in the control of metabolism genes, we tested whether PGC-1α could be recruited at promoter sites targeted by ERRα using chromatin immunoprecipitation (ChIP) experiments. ERBB2/Neu-induced breast cancer cells with increased expression of PGC-1α (α-1.1) displayed two-fold enrichment or greater for PGC-1α and ERRα at the promoter of glutamine metabolism genes (Figure 1C), supporting the tight association between the ERRs and PGC-1 s in the control of metabolic genes. Finally, we confirmed that the PGC-1α-mediated increase in expression of glutamine metabolism genes resulted in approximately two-fold or greater increase in protein levels (Figure 1D,E). This coordinated augmentation in mRNA and protein expression, in addition to corresponding enrichments observed in ChIP analyses, demonstrates that PGC-1α and ERRα exert a concerted control of glutamine metabolism genes; such pathway control has been shown to have strong biological relevance.

17To demonstrate the biological significance of the changes in glutamine gene expression, we first quantified relative glutamine uptake in ERBB2/Neu-induced breast cancer cells with increased expression of PGC-1α (α-1.1) and control cells. The α-1.1 cells showed a 25% increase in glutamine uptake both in normoxia and hypoxia compared with control (Figure 2A). This increase is similar in magnitude to that elicited by oncogenic levels of Myc.

18To gain insight on the flux of glutamine upon increased PGC-1α expression, α-1.1 and control cells were incubated with [U-13C]-glutamine labelled at all five carbons (Figure 2B). α-ketoglutarate m + 5 is generated from glutamine m + 5 ([U-13C]-glutamine) and serves as an entry point in the CAC, where it can either generate isocitrate m + 5 by reductive carboxylation (reverse flux) or succinate m + 4 by oxidative decarboxylation (forward flux) (Figure 2B). Reverse CAC flux leads to enrichments in citrate m + 5, malate m + 3, and fumarate m + 3, and forward CAC leads to the enrichment of m + 4 metabolites (Figure 2B). Citrate m + 4 and m + 5 can be used for lipid biosynthesis, but citrate generated specifically through reverse CAC (m + 5) plays a central role in glutamine-derived lipogenesis in cancer cells (Figure 2B).

19We performed glutamine labeling experiments in normoxia and hypoxia given that reductive carboxylation of glutamine is particularly active in hypoxia. In normoxia, α-1.1 cells displayed increased entry of glutamine in the CAC as illustrated by elevated amount of m + 5 α-ketoglutarate compared with control (Figure 2C). The α-ketoglutarate can then enter both the forward and reverse CAC reactions. The α-1.1 cells displayed higher forward and reverse CAC activity compared with controls (Figure 2C-F). The elevated forward flux can be seen from the higher amount of m + 4 citrate, malate, and fumarate, while the elevated reverse flux can be observed in the higher amount of m + 5 citrate as well as m + 3 malate and fumarate (Figure 2D-F) [see Additional file 4: Figure S2A]. The reverse flux represented a small fraction of the total CAC flux in normoxia. In hypoxia, there was an increase in the entry of glutamine in the CAC at α-ketoglutarate both for the control and α-1.1 cells compared with normoxia (Figure 2C). In general, hypoxia caused an increase in the reverse flux for both the control and α-1.1 cells (Figure 2D-F) [see Additional file 4: Figure S2A], but the effects were more pronounced in the α-1.1 cells and can be best appreciated when examining the labeling on citrate (Figure 2D). Indeed, α-1.1 cells displayed a large m + 5 labeling on citrate in hypoxia and devoted a larger fraction of their CAC flux for reverse reactions than control cells in hypoxia (Figure 2D). Together, these metabolomics experiments illustrate that PGC-1α can stimulate glutamine-mediated forward and reverse CAC fluxes in normoxia and hypoxia.

20Next, we investigated whether ERRα could also regulate glutamine flux through the CAC given that it can be recruited with PGC-1α to the promoter of glutamine metabolism genes. To do so, we used the ERRα pharmacological inhibitor C29 in the ERBB2+ human breast cancer cell lines SK-BR-3 and BT-474. Treatment of SK-BR-3 cells with C29 both in normoxia and hypoxia decreased the expression of glutamine metabolism and CAC genes (Figure 3A). This is supported by a concomitant decrease in glutamine metabolism protein levels [see Additional file 5: Figure S3]. In agreement with these changes, SK-BR-3 cells treated with C29 displayed a decrease in forward CAC flux, most notably in hypoxia, illustrated by reduced fraction of m + 4 citrate, malate, and fumarate (Figure 3C). The reverse CAC flux was significantly decreased in the presence of C29 in normoxia and hypoxia, demonstrated by a reduced fraction of m + 5 α-ketoglutarate and citrate as well as m + 3 malate and fumarate (Figure 3C). In BT-474 cells, C29 decreased the expression of glutamine metabolism and CAC genes and also reduced forward and reverse CAC fluxes in hypoxia (Figure 3B,C). The lack of effect of C29 in BT-474 cells in normoxia is most likely explained by the fact that these cells are not as dependent on ERRα activity as SK-BR-3 cells under standard conditions. Together, the metabolite flux experiments performed with PGC-1α and ERRα illustrate that they both play a central role in controlling glutamine-mediated forward and reverse CAC fluxes in ERBB2+ breast cancer cells. It is important to note that we did not detect glutamine-derived lactate in all ERBB2+ cells we tested [see Additional file 4: Figure S2B-D].

21One central function of glutamine metabolism is to replenish the CAC of intermediates that are used for biosynthetic processes, notably citrate that is used in lipid synthesis. Indeed, glutamine has been shown to play a central role in lipid biosynthesis in cancer cells. We measured the impact of PGC-1α on the expression of key genes involved in lipid biosynthesis. ERBB2/Neu-induced breast cancer cells with increased expression of PGC-1α (α-1.1 and α-1.2) displayed an elevated expression of lipogenic genes compared with control cells (Ctl-1; Figure 4A,B). In order to determine the involvement of PGC-1α in glutamine-mediated lipid biosynthesis, we incubated α-1.1 cells and control with trace amounts of [U-14C]-glutamine and monitored 14C incorporation in whole lipid extracts. The α-1.1 cells displayed a significant (48%) increase in 14C incorporation compared with controls in normoxia, as well as a robust (63%) increase in 14C incorporation compared with controls in hypoxia (Figure 4C). As expected, there was an important increase in glutamine-mediated lipid biosynthesis in hypoxia in both control and α-1.1 cells (Figure 4C). Also, ERRα is an important regulator of glutamine-mediated lipogenesis as BT-474 cells treated with C29 elicited a significant (62%) decrease in [U-14C]-glutamine incorporation into lipids in hypoxia [see Additional file 6: Figure S4A]. These data illustrate that PGC-1α and ERRα are key regulators of glutamine-mediated lipid biosynthesis.

22Next, we tested the impact of PGC-1α on the sensitivity of ERBB2/Neu-induced breast cancer cells to glutamine deprivation, low glucose, and hypoxia, in order to mimic the environment of solid growing tumors. The growth of both ERBB2/Neu-induced breast cancer cells with increased expression of PGC-1α (α-1.1) and controls was ablated in the absence of glutamine (Figure 4D), illustrating that ERBB2/Neu-induced breast cancer cells require glutamine for their growth. A similar dependency on glutamine was observed with human ERBB2+ breast cancer cell lines [see Additional file 6: Figure S4B,C]. We then tested the impact of PGC-1α on cellular proliferation in low glucose and hypoxia, two conditions that can favor glutamine usage (Figure 4E). The proliferation of α-1.1 cells was less sensitive to a combination of low glucose and hypoxia than controls as indicated by their smaller reduction in proliferation (Figure 4E). Together, these data demonstrate that glutamine is a key metabolite for the growth of ERBB2/Neu-induced breast cancer cells, and that the PGC-1α-mediated increase in glutamine metabolism is associated with protection against low glucose conditions and hypoxia, two conditions that can favor glutamine metabolism.

23We examined the expression of PGC-1α (PPARGC1A) and glutamine enzymes in 732 breast cancer patients from The Cancer Genome Atlas (TCGA) to determine their expression pattern across breast cancer subtypes and correlation with survival. PPARGC1A expression was highest in basal and ERBB2-enriched breast cancer subtypes, which have poorer prognosis than luminal A and B subtypes (Figure 5A). The expression of glutamine enzymes (glutamine cluster) was positively correlated with PPARGC1A expression in the basal and ERBB2-enriched breast cancer subtypes, in agreement with the fact that these subtypes have the highest expression of PPARGC1A (Figure 5B). Then, we determined if the expression of the glutamine cluster was correlated with clinical outcome. High expression of the glutamine cluster was associated with reduced survival in ERBB2-enriched and luminal B subtypes, and this association held true when looking at all breast cancer patients (Figure 5B,C). The fact that high expression of the glutamine cluster is associated with reduced survival in luminal B subtype but is not positively correlated with PPARGC1A expression suggests that, while glutamine metabolism is clinically relevant in this subtype, factors other than PGC-1α are driving the expression of this pathway. Overall, these data provide clinical relevance to the notion that PGC-1α is a central regulator of glutamine metabolism in ERBB2+ breast cancer cells.

24In this manuscript, we report that the PGC-1α/ERRα axis is a key regulator of glutamine metabolism in ERBB2+ breast cancer cells. We revealed that PGC-1α/ERRα bind the promoter of glutamine metabolism genes and regulate their expression, which ultimately controls the flux of glutamine in cells. This can contribute to increased lipid biosynthesis, especially under hypoxic conditions. Finally, we demonstrate that the PGC-1α-mediated control of glutamine metabolism has clinical relevance in ERBB2+ breast cancer patients.

25Even though the importance of glutamine metabolism in cancer cells is well appreciated, much less is known regarding central regulators of glutamine metabolic pathways. Major transcriptional regulators of glutamine metabolism identified so far are c-Myc and HIF-1. In this study, we reveal that PGC-1α/ERRα are central regulators of glutamine-mediated forward and reverse CAC fluxes. Interestingly, a recent report showed that glutamine reductive carboxylation for lipogenesis is particularly active in brown fat cells, the same cell type where PGC-1α was discovered and which display very high expression of this coactivator. In recent years, PGC-1s/ERRs have been involved in the control of the expression of numerous metabolic pathways in breast cancer cells. Both PGC-1α and ERRα have been shown to promote the growth of ERBB2+ tumors in vivo and to regulate the metabolism of this specific breast cancer subtype. Interestingly, increased expression of PGC-1α reduced the proliferation of ERBB2+ cells in vitro, but promoted tumor growth in vivo by promoting nutrient supply. We show here that ERBB2+ cells are dependent on glutamine for their growth as complete glutamine removal from cell culture severely compromised cell growth. Although the glucose dependence of ERBB2+ cells is well known, the role of glutamine in this system has been largely overlooked. A point supporting the importance of glutamine in ERBB2+ tumors is the observation that ERBB2 signaling elevates c-Myc expression, which is a central regulator of glutamine metabolism. Furthermore, it would not be surprising that the pro-growth effect of PGC-1α and ERRα in ERBB2+ tumors is mediated at least in part through their positive impact on glutamine metabolism. In support of this point, the expression of PGC-1α and that of the glutamine pathway are positively correlated in ERBB2+ breast cancer patients, and high expression of glutamine metabolism enzymes is associated with shorter survival. It is important to emphasize that we do not exclude the possibility that PGC-1α may associate with other transcription factors, in addition to ERRα, to control the expression of glutamine metabolism genes.

26In this paper, we reveal that the PGC-1α/ERRα axis is a central regulator of glutamine metabolism in ERBB2+ breast cancer. PGC-1α/ERRα likely regulate similar metabolic programs in diverse cancer types as the metabolic targets of PGC-1α/ERRα in various normal tissues and cancer cells/tumors are similar. The positive or negative impact of PGC-1α/ERRα on tumor growth will depend on the specific metabolic alterations of each tumor. There is significant effort currently underway to associate metabolic signatures to specific oncogenes. Furthermore, it is important to appreciate that the PGC-1α/ERRα axis also regulates the expression of genes not directly associated with metabolism; these genes may be equally or even more important in defining the roles of PGC-1α/ERRα in various cancer types. For example, ERRα regulates the expression of genes located in the ERBB2 amplicon that are involved in signal transduction, cell migration and invasion, as well as tamoxifen resistance. Overall, much remains to be discovered regarding the importance of PGC-1α/ERRα in contributing to the altered metabolism of various cancer types and to link this new metabolic state with clinical outcome. Also, the importance of PGC-1α/ERRα target genes involved in other hallmarks of cancer cells such as migration/invasion, apoptosis, and angiogenesis needs to be established. That being said, the identification of PGC-1α/ERRα as central regulators of glutamine metabolism may open novel therapeutic avenues for tumors showing strong dependence on this metabolic pathway.

27CAC: Citric acid cycle; ChIP: Chromatin immunoprecipitation; RPKM: Reads per kilobase of exon model per million mapped reads; TCGA: The Cancer Genome Atlas.

28The authors declare that they have no competing interests.

29JSP, RGJ, VG, SM and SPG designed the experiments; SM and SPG performed qPCR, immunoblotting, and protein quantification; GD performed ChIP analyses; SPG performed GC/MS and glutamine uptake experiments, with technical assistance and counseling by DA; SPG and SM performed mass isotopomer distribution analysis; SPG and DJP performed siRNA experiments; SM and BF performed 14C-lipid incorporation assays; SM performed proliferation experiments; AW and KH performed preliminary metabolomics experiments; UDA performed bioinformatics analyses; JSP, SM and SPG wrote the paper. All authors read and approved the final manuscript.

30Co-first authors: Shawn McGuirk and Simon-Pierre Gravel.